Datasheet

3−55
3.10 Output Crossbar Mixers
The TAS3103A has three serial output ports—SDOUT1, SDOUT2, and SDOUT3. Each serial output port is assigned
two processing nodes within the TAS3103A. One of the two nodes sources the left stereo data sample, and the other
node sources the right stereo data sample. Figure 3−35 shows the assignment of these internal nodes to the serial
output ports. Two cases are shown in Figure 3−35—discrete mode and TDM mode. The discrete mode connections
are straightforward, but the TDM connections are considerably more involved in order to support the different
one-chip and two-chip TDM modes. See Input and Output Serial Audio Port (SAPs), Section 2.1, for more discussion
on the TDM modes.
The purpose of the output crossbar is to give each of the three monaural channels in the TAS3103A access to any
of the six internal processing nodes (U, V, W, X, Y, and Z) that supply data to the three serial output ports. This flexibility
in the routing of the monaural channel outputs to the serial output ports, coupled with the flexibility in the routing of
the serial input ports to the monaural channels, fully decouples the input data from the output data. A given process
flow and output data topology can be obtained from any ordering of data into the TAS3103A.
Figure 3−36 shows the output crossbar mixer topology. Each monaural channel feeds six mixers. The six mixers, in
turn, feed the six output nodes U, V, W, X, Y, and Z. A given monaural channel can thus be connected to either the
left or right side of SDOUT1, SDOUT2, and SDOUT3.
The mixers, although capable of performing boost (gain) and cut (attenuation) on the outgoing audio, are typically
used to facilitate on/off switching (a 5.23−format coefficient value of 0x0800000 turns the mixer on and a coefficient
value of 0x0000000 turns the mixer off). The audio data streams at the input to these mixers include dither, and any
boost or cut in the audio at this point affects the dither levels as well.
Node r in Figure 3−36 provides a means of outputting a post-processed sum of the audio on channel 1 and channel
2. This capability could be used to generate a center audio component from L and R components being processed
on channels 1 and 2. This would allow channel 3 to be a subwoofer channel. Node r could also be used to create a
subwoofer channel, assuming an active subwoofer with filtering capability is receiving the subwoofer output. This
option would then free channel 3 for center channel processing.